Bill Schertz wrote:
Everyone,
I discovered an error in my calculations that I presented at the
Rotary
Roundup on Saturday. I have enclosed the revised graphs and sample
calculation as a PDF file that everyone should be able to open.
Thanks to those that questioned my numbers -- the methodology was
correct, but my data entry wasn't so good.
please use the attached figures in liu of the ones in the handout.
Bill Schertz
KIS Cruiser # 4045
Bill will be sending us a complete PDF of his entire excellent talk
shortly I suspect.
It is a unified cooling theory useful for predicting the
specifications of
a complete
water side rotary engine cooling system.
Paul Lamar
Paul,
I follow the maths OK but I am confused on the dT 10 deg-f.
Is that for every 10 degrees above some standard ( or known)temperature
or
do you have to measure the temp before it goes into the block.
Also what is the flow rate of the Mazda 13b water pump.
George (down under)
I'll let Bill answer that.
Paul Lamar
A fluid carries heat by having its temperature raised. If you have a very
small rate of flow, then you need a large temperature rise (or drop) to
carry the heat from engine to radiator.
[The higher the water flow in the block the less the temp rise in the water.
Also as an aside high water flow causes turbulence which enhances the
heat transfer from the metal block to the water. PL]
The water may have a 20 to 30 degree rise across the engine (drop across the
radiator), and the air may have a larger rise across the radiator to take
the heat away.
I will send paul the PDF tomorrow.
Bill Schertz
KIS Cruiser # 4045
What Bill is saying here is the water temperature increase through the block must be the same
as the water temperature decrease across the rad or the engine will over heat.
Obvious when you think about it.
Now we are left with the problem of how to lower the water temperature through the
rad so these two water delta T's equal each other. The water will heat the metal
in the first part of the radiator to near the water temp initially.
Now the third delta T comes into play. Namely the difference in temperature between
the rad metal (220F) and the air flowing through the rad (110 F worst case Arizona).
Since the rate of heat transfer is more or less directly proportional to this temperature
difference among other things it behooves us to keep the air as cool as possible.
For low aerodynamic drag we want to keep the air pressure drop to a minimum. That means a large
area thin rad. For high radiator effectiveness we also want a thin rad. The reasoning behind
this is as the air flows through the rad it warms up so the delta T decrease decreasing
the rate of heat transfer. The rule of thumb is the first half of the rad does 3/4 of
the cooling while the second half does one quarter. The third factor is we have a modest
dynamic pressure to force air through the rad to work with at slow climb speeds.
At 100 MPH the maximum pressure we can expect is only about .18 psi. Therefore
ideal heat exchanger would be zero thickness and infinitely large.
The other major factor is the total area of metal in contact with both the water and the air. Generally speaking the heat transfer from water to the metal is higher than the heat
transfer from metal to air. That is the reason for the fins between the tubes in a
typical radiator. Turbulence in the air in contact with the fins improves the heat
transfer from the metal to the air. That is achieved by punching tiny louvers in the
fins. Unfortunately that increases the air pressure drop through the rad increasing
the aerodynamic drag. As you can see there are huge tradeoffs and compromises.
Now we have to fit this rad under the cowl. What seems to be needed is a rad of
around 600 to 700 cubic inches. Since we are limited to rad cores of about 19 inches
long and about 16 inches wide we need a core about 2 inches thick. Not ideal
but practical to get inside the cowl along with the engine.
The next thing to worry about is equally distributing the air flow over
the face of the rad. Kuchemann & Weber and Kays and London both addressed
this issue. K&M came up with a relatively long trumpet shaped duct and vertical rad
which due to its length is impractical for most of our airplanes. It was used on
the final experimental model of the P51. K&L on the other hand mounted the rad
horizontally and used a simple wedge shaped duct to feed air through the rad forcing
it to make a 90 degree turn. This allowed mounting the rad under the cowl
without increasing the frontal area of the fuselage. The P51 style scoop on
the other hand increases the frontal area of the fuselage. The frontal area is
the maximum cross sectional area of the fuselage no matter where it occurs along
it's length. All other things being equal the drag of the fuselage is proportional
to its frontal area. NACA style flush ducts don't work when there is a pressure
build up in the opening as occurs when they are used with heat exchangers.
The air merely flows on by.
The jury is still out on which of these two configurations
is the lowest overall drag. Also the P51 was designed from the outset with the rad in
the aft fuselage. It is just not practical to cut large holes in current kit
planes to install the rad in the aft fuselage.
Paul Lamar
The Rotary Engine NewsLetter. Powered by Linux.
ACRE NL web site. http://home.earthlink.net/~rotaryeng/
Copyright 1998-2004 All world wide rights reserved.